robustSingleCell (version 0.1.1)

cluster.analysis: Distributed Clustering Analysis

Description

Perform clustering analysis for a range of hyperparameter (KNN Ratios) values and assess clustering quality relative to simulation analysis of shuffled data.

Usage

cluster.analysis(environment, knn.ratios = c(0.01, 0.05, 0.1),
  nShuffleRuns = 10, shuffledKNN = 10, loadPreviousKnn = T,
  rerun = F, deleteCache = F, mem = "4GB", time = "0:15:00",
  plot = T, local = F)

Arguments

environment

environment object

knn.ratios

range of KNN parameters to scan (corresponding to different resolutions)

nShuffleRuns

number of shuffled clustering analyses to perform per KNN threshold

shuffledKNN

number of closest KNN shuffled analyses to include in background clustering quality computation

loadPreviousKnn

whether to load previous analysis results

rerun

whether to rerun the analysis rather than load from cache

deleteCache

whether to delete cache files

mem

HPC memory

time

HPC time

plot

whether to plot the clustering qualities compared to shuffled

local

whether to run jobs locally rather than using distributed slurm system

Value

environment parameter containing clustering assignment and provisional cluster names

Examples

# NOT RUN {
LCMV1 <- setup_LCMV_example()
LCMV1 <- get.variable.genes(LCMV1, min.mean = 0.1, min.frac.cells = 0,
min.dispersion.scaled = 0.1)
LCMV1 <- PCA(LCMV1)
LCMV1 <- cluster.analysis(LCMV1)
# }